Espritmodel.com Telemetry Radio
Reply
Thread Tools
Old Jul 12, 2012, 11:51 PM
Registered User
Joined Oct 2011
37 Posts
title

the image from one of the patent pdf's says that it uses two 16 bit micro controllers are used . any thoughts ?
u.abhi19 is offline Find More Posts by u.abhi19
Reply With Quote
Sign up now
to remove ads between posts
Old Jul 13, 2012, 11:09 PM
Enjoying the sunshine
HappySundays's Avatar
Australia, NSW, Sydney
Joined Feb 2011
2,893 Posts
Quote:
Originally Posted by u.abhi19 View Post
built the structure for the 1st prototype .. finding really hard to get the control movements and their respective hex codes for my atmega 168pa .. used spdf-vtol configuration but the firmware is not optimum for the use of 8 servos .. open aero firmware is loading but its not arming .. tried everything possible .. please can any1 help me out?
Waiiit a minute...
From your pictures you have made your model quite differently to a normal ball drone.
If you look at the pictures the normal one has FOUR actuators at the TOP, then the motor, the FOUR actuators at the BOTTOM... no wonder you had trouble getting it to work with OpenAero...

If you look back at this post the RED actuators are on the top and the BLUE ones at the bottom. I thought it was pretty clear from my video...
HappySundays is offline Find More Posts by HappySundays
Reply With Quote
Old Jul 15, 2012, 10:39 PM
Registered User
Joined May 2012
2 Posts
Quote:
Originally Posted by u.abhi19 View Post
built the structure for the 1st prototype .. finding really hard to get the control movements and their respective hex codes for my atmega 168pa .. used spdf-vtol configuration but the firmware is not optimum for the use of 8 servos .. open aero firmware is loading but its not arming .. tried everything possible .. please can any1 help me out?

can you give me a clear view of the servo and the vane(the last picture)? I want to see how you put the servo and the vane. I'm interested in how your servo control the movements of the vane.

Thanks.
suvicky is offline Find More Posts by suvicky
Reply With Quote
Old Jul 23, 2012, 09:34 AM
Registered User
Joined Oct 2011
37 Posts
new design..

............
u.abhi19 is offline Find More Posts by u.abhi19
Last edited by u.abhi19; Jul 25, 2012 at 08:12 AM.
Reply With Quote
Old Jul 26, 2012, 05:47 AM
Registered User
Joined Sep 2008
25 Posts
I already submitted almost same design in a government research laboratory 5 years ago.

http://www.rcgroups.com/forums/showthread.php?t=933109

only a rough schematic here on forum, i once posted.
zed1234 is offline Find More Posts by zed1234
Reply With Quote
Old Aug 23, 2012, 09:41 AM
Registered User
Joined Feb 2012
21 Posts
Quote:
Originally Posted by HappySundays View Post
Yesterday I posted a video of the ball drone mixing in OpenAero. Today here's a video showing the gyro stabilisation completed for the ball drone configuration.
So basically all you need is a cheap HobbyKing V2 KK board and this and some fiddling and you should be able to have a controllable, stabilised ball drone.

I haven't updated the GUI for it but you can set it up via LCD as per usual. If there is sufficient interest I'll provide more details, code etc.

http://youtu.be/wWv5F1_Fm9U
tha


Thanks for the wonderful info. It seems like the direction of the vanes are different from the ones Sam has demonstrated in his video. The upper and lower vanes moves in opposite direction with respect to each other when performing pitch and roll. I'm confused now... Would someone enlighten me?
paperbert is offline Find More Posts by paperbert
Reply With Quote
Old Sep 25, 2012, 10:05 AM
Registered User
Joined Sep 2012
13 Posts
Hi Sam,

Im Tim from Singapore, and I am working and researching on the Japanese ball drone too.

I have been following your 'Taiwan Ball' development from this site: http://www.rcgroups.com/forums/showt...454102&page=17

Im having some trouble with the stability and control of the ball. I was wondering if you are able to help.

Basically, I have the same control settings as discussed in the forum, using 2 mixers for the top and bottom sets of flaps to control roll/pitch and yaw. The only difference is that instead of the piezo gyros, i am using Ardupilot Mega 1.0, and I have been trying all means to get it to hover but it seems like it is always either slipping (losing lift) or oscillates due to recession even though the yaw control is fully activated.

Any advice or stages to help me achieve stability? I understand that the first step is to try to even get the ball to hover without the use of the gyros, but it cant seem to work.

Most PID settings see the ball moving in random slipping motions as shown below:
P=1.6, I=0, D=0.6 (0 min 46 sec)


However, this shows the best that I have achieved so far, but I have ran out of ideas on how I can further improve it.
P=1.6, I=0, D=0.5 (0 min 44 sec)


Anyone who is reading, feel free to give me some input as well

Thank you so much in advance.

Regards,
Tim
tim2chia88 is offline Find More Posts by tim2chia88
Last edited by tim2chia88; Sep 25, 2012 at 10:13 AM.
Reply With Quote
Old Sep 25, 2012, 11:04 AM
DIY Mania from Taiwan
Taiwan
Joined Aug 2011
1,900 Posts
Tim:you must have spent a lot of efforts for the drone,second vid looks nice
cut the weight,adjust the servos range,trim the PID and you are in the club.
congratulations
regards
Sam
samwei1950 is offline Find More Posts by samwei1950
Reply With Quote
Old Sep 25, 2012, 08:06 PM
Registered User
Joined May 2010
373 Posts
CG Location?

Hello Tim,

That is first rate construction work on your sphere.

Where is your CG location?

I'd like to suggest that you lower the CG a very small amount and test fly the craft, repeat this task several times until you see the outcome of an excessive low CG location, film each change in the lowering of the CG, then report back your findings.

After the CG test, I would suggest that you eliminate the D component and increase the I component, but not until you make the CG test!

Regards, and keep-up the great work.

Kelly
corocopter is offline Find More Posts by corocopter
Reply With Quote
Old Sep 25, 2012, 08:51 PM
Registered User
Joined May 2010
373 Posts
Tim,

I'm very sure the original Jap Drone uses 3 axis mems gyros and 3 axis linear accelerometers in order to have the rock solid control we see in his videos.

Does your board have these components?

Sam did very well with cheap gyros, but his craft does not match the performance of the original ship, but I'm still very impressed with how well Sam was able to get such a low cost craft to hover.

You may need to increase the size of your lower control vanes, it is hard to see exactly how large they in comparison to Sams.

Are you using the Canarderon setup? That is to say, upper vanes moving opposite the lower vanes for pitch and roll. Which vanes are you using for yaw control?

If yaw is not holding, you maybe under powered in two areas

1, insufficient size of the control surfaces (vanes).
2, insufficient deflection of the vanes, which require some off-set at all times to compensate for motor torque.

The more details you give me the more I can help.

Kelly
corocopter is offline Find More Posts by corocopter
Reply With Quote
Old Sep 25, 2012, 09:07 PM
Registered User
Joined May 2010
373 Posts
Tim,

I would also suggest that you paint the center-line vertical rib black in color to allow you to visualize the yaw axis better.

Sam used a ping pong ball to good affect too.

Regards,

Kelly
corocopter is offline Find More Posts by corocopter
Reply With Quote
Old Sep 25, 2012, 09:41 PM
Registered User
Joined Sep 2012
13 Posts
Hi Corocopter,

Thank you for your comments! I have learnt so much from reading all your posts in this thread.

Let me explain a little about how i came about at the PID values

1) From the 'bench test', I tested using P=1, I=0, D=0. After a few rounds of tests, I found that P=1.6, I=0, D=0 dampens the number of oscillations, without overdamping or underdamping.
(Im not sure if the results from the test was fair, do comment on the experimental test carried out as seen from the video )
Bench Test P=1, I=0, D=0 (0 min 30 sec)

(also help yourself to at all the other tests which I have done, in my channel)

2) I decided to test fly the ball starting with P=1.6, I=0, D=0, increasing D to 0.1 gave similar results. Only after increase D=0.5, i arrived at the best values so far at P=1.6, I=0, D=0.5
Just to wondering if you are familiar with PID, I would like to check if my concept of those settings are correct
Basically an increase in P increases the gain, and thus the amount of flap angle, during correction by the gyro.
Increasing I will decrease the steady state error, but decrease stability.
Finally, increasing D will increase stability, but reduce steady state error.
Im not too sure how steady state error can be observed. But I could tell when there is instability from the oscillatory effect.
Therefore, knowing P and D goes hand in hand, we decided to work on P and D alone. There are videos where we had some I input, but it wasnt too ideal.

Honestly, I could not really come to a conclusion the effects of changes to PID settings, maybe Im not sure what Im expected to observe when I make changes to either of the settings.

Also, another point to note, I do understand that we have to test fly without gyros to observe static balance in the air before even moving on to test fly with the gyros activated. Do I have to ensure that the ball can hover for about 5 secs without effects of yaw before fine tuning the controls with the gyro?

Basically after doing all the PID testings, it seems that I have no luck in getting those 'magic numbers', and I suspect that it could be do to a problem with the CG, whether it is even in line with the vertical plane, because I tend to get precession from some of the test flights observed.
Currently my CG is below the meridian by 5cm. Ill try to play around with the height of CG and get back to you

Im jus writing whatever that has come to my mind from the last few weeks of research, sorry if everything seems so incoherent :P

Any advice or comments from anyone would be good advice for me.Thanks in advance and hope to hear from you soon

Tim
tim2chia88 is offline Find More Posts by tim2chia88
Reply With Quote
Old Sep 25, 2012, 10:08 PM
Registered User
Joined Sep 2012
13 Posts
Quote:
Originally Posted by corocopter View Post
Tim,

I'm very sure the original Jap Drone uses 3 axis mems gyros and 3 axis linear accelerometers in order to have the rock solid control we see in his videos.

Does your board have these components?

Sam did very well with cheap gyros, but his craft does not match the performance of the original ship, but I'm still very impressed with how well Sam was able to get such a low cost craft to hover.

You may need to increase the size of your lower control vanes, it is hard to see exactly how large they in comparison to Sams.

Are you using the Canarderon setup? That is to say, upper vanes moving opposite the lower vanes for pitch and roll. Which vanes are you using for yaw control?

If yaw is not holding, you maybe under powered in two areas

1, insufficient size of the control surfaces (vanes).
2, insufficient deflection of the vanes, which require some off-set at all times to compensate for motor torque.

The more details you give me the more I can help.

Kelly
Yes I am impressed with sam's taiwan ball, he seems to have not that much trouble with stability even on his very first post. Im wondering how he did that.

1) yes im using the Ardupilot Mega 1.0 which comes with 3axis control.

2) "Are you using the Canarderon setup? That is to say, upper vanes moving opposite the lower vanes for pitch and roll. Which vanes are you using for yaw control?"

Im not sure if i get you right, r you saying if i wanna pitch/roll left, my top vanes have to move opposite the lower vanes? because now, when i wanna pitch/roll right, my top vanes will move left, and 4 of the bottom vanes will go left as well. which is shown in the picture attached
tim2chia88 is offline Find More Posts by tim2chia88
Reply With Quote
Old Sep 26, 2012, 12:46 PM
Registered User
Joined May 2010
373 Posts
Vane defection error:

Quote:
Originally Posted by tim2chia88 View Post

Kelly wrote;
2) "Are you using the Canarderon setup? That is to say, upper vanes moving opposite the lower vanes for pitch and roll. Which vanes are you using for yaw control?"

Tim wrote:
Im not sure if i get you right, r you saying if i wanna pitch/roll left, my top vanes have to move opposite the lower vanes? because now, when i wanna pitch/roll right, my top vanes will move left, and 4 of the bottom vanes will go left as well. which is shown in the picture attached
Tim,

Good news, I see your problem clearly, your comment above is correct. Your current control configuration of vane deflection needs to be changed in one aspect.

First let me comment on the things I see that are correct or nominal for the time being.

Yaw control (lower vanes) deflection is correct.

CG is okay for now.

PID is okay for now.

Control vane size or "Volume" as termed in aeronautics is okay for now
.
Spoiler locations are okay, but they need weights (small steel washers) attached to fine tune each spoiler so that they will deploy downward in translational flight. Save this for latter, after you have a stable hover.

What is wrong

Pitch and roll defection of the upper vanes is reversed

I repeat this once more and add details>>>>>

Your control defection of the upper vanes (orange color in the diagram) is wrong. You must reverse their direction. Don't feel bad, you copied this diagram from post #289 by HappySundy, it was not his error, he was asked to make a program by another poster that made the mistake, so many thanks for the efforts HappySunday.

Canarderons work much like a 4 wheel STEERING car. Imagine turning right while looking downward on a 4 wheel steering car you will notice the front wheels turn to the right and the aft wheel turn to the left, thus the car rotates around the center point of the chassis and hopefully, center of mass. This allows for a very small turning radius and very rapid response. The same physics apply to the spheres. Please return to my description of this matter as I posted a detailed explanation of this concept to Sam on this thread. This is how the original Jap Sphere was configured, and it must be configured in this manner to fly.

Your craft will fly great after this program change.

We can go over many details to fine tune the sphere after you have a solid, well controlled hover. You're definitely on the right track and nearly ready to go back to PID matters, for now your PID looks okay. The basic problem in upper control vanes deflection, they are reversed. Just fix that one thing for now and test fly. Please post another Vid of your test flight.

Best regards,

Kelly
corocopter is offline Find More Posts by corocopter
Reply With Quote
Old Sep 26, 2012, 07:17 PM
Registered User
Joined May 2010
373 Posts
Quote:
Originally Posted by tim2chia88 View Post
Just to wondering if you are familiar with PID, I would like to check if my concept of those settings are correct
Basically an increase in P increases the gain, and thus the amount of flap angle, during correction by the gyro.
Increasing I will decrease the steady state error, but decrease stability.
Finally, increasing D will increase stability, but reduce steady state error.
Im not too sure how steady state error can be observed. But I could tell when there is instability from the oscillatory effect.
Therefore, knowing P and D goes hand in hand, we decided to work on P and D alone. There are videos where we had some I input, but it wasnt too ideal.

Honestly, I could not really come to a conclusion the effects of changes to PID settings, maybe Im not sure what Im expected to observe when I make changes to either of the settings.

Also, another point to note, I do understand that we have to test fly without gyros to observe static balance in the air before even moving on to test fly with the gyros activated. Do I have to ensure that the ball can hover for about 5 secs without effects of yaw before fine tuning the controls with the gyro?

Basically after doing all the PID testings, it seems that I have no luck in getting those 'magic numbers', and I suspect that it could be do to a problem with the CG, whether it is even in line with the vertical plane, because I tend to get precession from some of the test flights observed.
Currently my CG is below the meridian by 5cm. Ill try to play around with the height of CG and get back to you

Im jus writing whatever that has come to my mind from the last few weeks of research, sorry if everything seems so incoherent :P


Tim
In response to your questions/comments:

I understand the very basics of PID, it would be nice to have an expert from the quad-copter crowd help us out here.

I don't understand what the term "steady state" refers to and how it applies to the I term, maybe you can add some thoughts to this idea?

My PID understanding is more related to the calculus concepts and I have not had the pleasure of owning a stabilization device that I could adjust the PID, only some low cost gyros. So I'm of little help giving you much feed back on your You-Tube videos. One thing you can feel good about, is the fact that your controls here setup wrong, yet you where able to maintain a degree of stability under these condition that is very impressive! You'll get a much better understanding of the interactions of PID when you have the Canarderons working.

I'd suggest you leave the CG alone for now, that sound just about right from Sams feed back.

The craft must balance vertically, ie. if you hang the craft from the propeller shaft (not the external outer cage) it must hang exactly vertical. I showed how to use a bubble level to make this measurement on this thread. This is extremely important on any VTOL aircraft and this fact is almost always overlooked when 3D aircraft are balanced. It shows up in hover and can be viewed by the fact that the fuselage is not vertical when in a hover, additionally a conventional aircraft requires the same amount of vertical stabilizer both above and below the fuselage to deal with the swirling vortex of air created by the prop, it should be a mirrored image on top and on the bottoms, this look unsightly and makes for very long landing gear legs, but it makes hovering super easy!!!

Your comment are very coherent and are of great value, keep up the great feedback. There is much for both of use to learn from your experiments

Kelly
corocopter is offline Find More Posts by corocopter
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Mini-HowTo RC Controlled AR.Drone w/o WiFi! nosaari Multirotor Talk 116 Mar 14, 2014 02:23 PM
Discussion New AR Drone (i-drone) light mod bungle68 Multirotor Talk 3 Apr 13, 2011 05:30 PM
Discussion AR Parrot Drone For Sale Matt Gunn Atlanta Area RC 4 Mar 16, 2011 07:41 AM
Found Parrot AR. DRONE iceman77 Aircraft - Electric - Helis (FS/W) 3 Mar 15, 2011 01:51 PM
Discussion Piloting Drone From iDevice BUT takeoff/landing from a computer chope1 Multirotor Talk 0 Mar 07, 2011 10:12 AM